Activity
From 04/06/2016 to 05/05/2016
05/05/2016
- 11:22 PM Bug #15755 (Resolved): rbd-mirror crashed with assertion failure during the sync in-progress
- Description of problem:
-----------------------
Hosted a VM using a Qemu client on a 10GB image created on primary ... - 09:56 PM Bug #15750: Journal corruption detected during rbd-mirror replay
- Last entry in the journal was a 16MB IO event -- which ballooned the final size of the journal object to ~44MB. Howe...
- 09:42 PM Bug #15750 (Resolved): Journal corruption detected during rbd-mirror replay
- Mirroring was enabled and OS installation started on VM:...
- 09:25 PM Bug #15749 (Closed): rbd journal inspect / export tools should start at specific client commit po...
- Given a large journal that has been trimmed, without starting at a specific (valid) commit position, it's likely you ...
- 05:57 PM Documentation #15746 (Resolved): Update mirroring guide to include pool/image status commands
- 03:04 PM Feature #15721 (Fix Under Review): rbd map: more helpful error messages
- PR: https://github.com/ceph/ceph/pull/8944
- 01:28 PM Bug #15028: Deep-flatten incorrectly updates fast-diff stats
- 01:16 PM Bug #15248: Flag is showing "flags: object map invalid", after enabling object map on an exclusiv...
- @Rohith: yes, fast-diff follows the same rationale. The fast-diff feature is an extension to the object-map that use...
- 10:58 AM Bug #15248: Flag is showing "flags: object map invalid", after enabling object map on an exclusiv...
- Jason Dillaman wrote:
> This is expected behavior -- when you enable object map on a pre-existing image, the object ... - 07:51 AM Backport #15743 (Resolved): jewel: Demote image format 1 error message on open to information mes...
- https://github.com/ceph/ceph/pull/9003
05/04/2016
- 10:11 PM Subtask #15688 (In Progress): Stub API integration between librbd and the new client-side cache
- 10:10 PM Feature #15656 (New): Initial python APIs to support mirroring
- 09:16 PM Subtask #15144 (Pending Backport): rbd CLI to retrieve rbd mirror state for a pool / specific image
- 09:06 PM Bug #15642 (Fix Under Review): "librbd::object_map::RefreshRequest: object map smaller than curre...
- *PR*: https://github.com/ceph/ceph/pull/8937
- 02:53 PM Bug #15642 (In Progress): "librbd::object_map::RefreshRequest: object map smaller than current ob...
- 08:46 PM Bug #15577 (Pending Backport): Demote image format 1 error message on open to information message
- 05:37 PM Bug #15577 (Fix Under Review): Demote image format 1 error message on open to information message
- *PR*: https://github.com/ceph/ceph/pull/8932
- 05:33 PM Bug #15577 (In Progress): Demote image format 1 error message on open to information message
- 06:32 PM Backport #15736 (Resolved): jewel: rbd-mirror admin socket commands to start/stop/restart
- https://github.com/ceph/ceph/pull/9010
- 04:35 PM Feature #15721: rbd map: more helpful error messages
- Josh Durgin wrote:
> We don't have much in the way of a feedback channel from the kernel, but we could at least poin... - 01:04 PM Feature #15721 (In Progress): rbd map: more helpful error messages
- 12:16 PM Feature #15721: rbd map: more helpful error messages
- @Venky: all yours -- I don't have the ability to assign you until you get added to the developer role.
- 06:45 AM Feature #15721: rbd map: more helpful error messages
- Please assign this to me. I'll work on it.
- 02:52 PM Subtask #15718 (Pending Backport): rbd-mirror admin socket commands to start/stop/restart
- 04:25 AM Subtask #15110 (Fix Under Review): Image sync should sync full snapshot state
- *PR*: https://github.com/ceph/ceph/pull/8921
05/03/2016
- 08:14 PM Feature #15721 (Resolved): rbd map: more helpful error messages
- We don't have much in the way of a feedback channel from the kernel, but we could at least point users to dmesg. We c...
- 06:17 PM Subtask #15710 (Closed): The exclusive lock policy should support skipping the journal
- 03:25 PM Subtask #15710: The exclusive lock policy should support skipping the journal
- A workaround exists right now (unset the feature bit in memory) -- probably not worth the added complexity.
- 01:56 PM Subtask #15710 (Closed): The exclusive lock policy should support skipping the journal
- The rbd-mirror daemon, when propagating deletes, will need to be able to update the local image (which should have a ...
- 06:16 PM Subtask #14420 (Resolved): New RBD notification pool object for recording mirroring errors
- 06:04 PM Subtask #15718: rbd-mirror admin socket commands to start/stop/restart
- *PR*: https://github.com/ceph/ceph/pull/8809
- 06:04 PM Subtask #15718 (Resolved): rbd-mirror admin socket commands to start/stop/restart
- 05:52 PM Tasks #15611: Need fsx test against the rbd-nbd device within teuthology
- Requires issue #15715 to be fixed before the fsx test can reliably run.
- 05:51 PM Backport #15716 (Resolved): jewel: RBD image mounted by command "rbd-nbd" the status is read-only.
- https://github.com/ceph/ceph/pull/9005
- 05:47 PM Bug #15604 (Pending Backport): RBD image mounted by command "rbd-nbd" the status is read-only.
- 12:09 PM Bug #15604 (Fix Under Review): RBD image mounted by command "rbd-nbd" the status is read-only.
- *PR*: https://github.com/ceph/ceph/pull/8775
- 05:45 PM Bug #15715 (Resolved): rbd-nbd does not properly handle resize notifications
- The daemon attempts to listen for image header notifications to detect when an image has been resized. However, if t...
- 05:35 PM Backport #15713 (Resolved): jewel: librbd: resize hang - use after free
- https://github.com/ceph/ceph/pull/9009
- 05:35 PM Backport #15712 (Resolved): jewel: Exclusive lock not properly released if journal non-primary
- https://github.com/ceph/ceph/pull/9007
- 05:20 PM Bug #15709 (Pending Backport): Exclusive lock not properly released if journal non-primary
- 02:18 PM Bug #15709 (Fix Under Review): Exclusive lock not properly released if journal non-primary
- *PR*: https://github.com/ceph/ceph/pull/8910
- 01:52 PM Bug #15709 (Resolved): Exclusive lock not properly released if journal non-primary
- If exclusive lock fails to be acquired because the journal is non-primary (replay cannot proceed), the lock is not re...
- 05:09 PM Bug #15690 (Pending Backport): librbd: resize hang - use after free
- 11:43 AM Bug #15690 (Fix Under Review): librbd: resize hang - use after free
- *PR*: https://github.com/ceph/ceph/pull/8907
- 03:10 PM Bug #15561 (Need More Info): RBD image can be listed but not opened
- It sounds like the image may have been partially removed or only partially created - you should be able to complete r...
- 02:24 PM Bug #15456 (Resolved): rbd bench-write shouldn't crash upon write failure
- 02:21 PM Feature #14540 (Pending Backport): rbd disk-usage CLI command should support calculating full ima...
- 12:43 PM Backport #15708 (Resolved): jewel: Seeing rollback and revert option for snapshot, which seems to...
- https://github.com/ceph/ceph/pull/9004
- 12:27 PM Feature #15706 (Resolved): Optionally limit the maximum number of snapshots for an image
- 11:49 AM Bug #15521 (Pending Backport): Seeing rollback and revert option for snapshot, which seems to be ...
- 11:31 AM Bug #15693 (Duplicate): common/lockdep.cc: 330: FAILED assert(id == -1) in TestLibRBD.TestCreateL...
- Same issue as #15690
- 05:33 AM Bug #15693 (Duplicate): common/lockdep.cc: 330: FAILED assert(id == -1) in TestLibRBD.TestCreateL...
found in a jenkins "make check" run . but the change being tested is not relevant: it was a cmake change:
<pre...- 08:52 AM Backport #15698 (Resolved): jewel: rbd-mirror: lockdep error in TestImageReplayer.Bootstrap
- https://github.com/ceph/ceph/pull/9008
- 08:52 AM Backport #15697 (Resolved): jewel: python: clone operation will fail if config overridden with "r...
- https://github.com/ceph/ceph/pull/8972
05/02/2016
- 11:40 PM Bug #15690: librbd: resize hang - use after free
- I think this hit https://jenkins.ceph.com/job/ceph-pull-requests/5026/console
- 11:24 PM Bug #15690 (Resolved): librbd: resize hang - use after free
- From running RBD_FEATURES=1 ./unittest_librbd...
- 11:36 PM Bug #15685 (Pending Backport): python: clone operation will fail if config overridden with "rbd d...
- 07:09 PM Bug #15685 (Fix Under Review): python: clone operation will fail if config overridden with "rbd d...
- *PR*: https://github.com/ceph/ceph/pull/8883
- 05:56 PM Bug #15685 (In Progress): python: clone operation will fail if config overridden with "rbd defaul...
- 05:19 PM Bug #15685 (Resolved): python: clone operation will fail if config overridden with "rbd default f...
- The "clone" call should implicitly override the default format since cloning in implicitly always using image format 2.
- 09:20 PM Bug #15521 (Fix Under Review): Seeing rollback and revert option for snapshot, which seems to be ...
- *PR*: https://github.com/ceph/ceph/pull/8698
- 09:15 PM Bug #15664 (Pending Backport): rbd-mirror: lockdep error in TestImageReplayer.Bootstrap
- 02:41 PM Bug #15664 (Fix Under Review): rbd-mirror: lockdep error in TestImageReplayer.Bootstrap
- *PR*: https://github.com/ceph/ceph/pull/8877
- 11:31 AM Bug #15664 (In Progress): rbd-mirror: lockdep error in TestImageReplayer.Bootstrap
- 06:09 PM Subtask #15688 (Resolved): Stub API integration between librbd and the new client-side cache
- 06:09 PM Feature #15687 (New): Client-side, disk-backed cache
- The initial iteration will support ordered writeback from a local disk-based (SSD/NVMe) cache. This will reduce the ...
- 05:46 PM Bug #15127 (Closed): fsx tests taking too long and possibly stuck
- Issue was seen only on that specific run, Closing.
- 02:42 PM Bug #15127: fsx tests taking too long and possibly stuck
- @Vasu: can we close this ticket?
- 06:47 AM Backport #15668 (In Progress): jewel: rbd disk-usage CLI command should support calculating full ...
- 06:44 AM Backport #15669 (In Progress): jewel: Errors encountered disabling object-map while flatten is in...
- 06:41 AM Backport #15666 (In Progress): jewel: rbd CLI to retrieve rbd mirror state for a pool / specific ...
- 06:33 AM Backport #15661 (In Progress): jewel: Parent image is closed twice if error encountered while ope...
- 06:30 AM Backport #15660 (In Progress): jewel: rbd-mirror can crash if start up is interrupted
- 06:25 AM Backport #15649 (In Progress): jewel: possible double-free of object map invalidation request upo...
04/29/2016
- 05:13 PM Subtask #15670 (Resolved): Delete image when a resync is requested
- The local rbd-mirror daemon process should handle the case where an image has been requested to be re-synced. This s...
- 05:10 PM Backport #15669 (Resolved): jewel: Errors encountered disabling object-map while flatten is in-pr...
- https://github.com/ceph/ceph/pull/8869
- 04:54 PM Backport #15668 (Resolved): jewel: rbd disk-usage CLI command should support calculating full ima...
- https://github.com/ceph/ceph/pull/8870
- 04:53 PM Bug #15572 (Pending Backport): Errors encountered disabling object-map while flatten is in-progress
- 04:15 PM Backport #15666 (Resolved): jewel: rbd CLI to retrieve rbd mirror state for a pool / specific image
- https://github.com/ceph/ceph/pull/8868
- 04:15 PM Subtask #15144: rbd CLI to retrieve rbd mirror state for a pool / specific image
- PR: https://github.com/ceph/ceph/pull/8442
- 04:02 PM Bug #15665 (Resolved): Live journal replay might skip entries from previous object set
- The last entry from a previously polled version of an object resulted in the object being trimmed from the replay set...
- 03:01 PM Bug #15664 (Resolved): rbd-mirror: lockdep error in TestImageReplayer.Bootstrap
- http://qa-proxy.ceph.com/teuthology/joshd-2016-04-28_23:45:43-rbd-wip-josh-testing---basic-mira/156104/teuthology.log
- 01:01 PM Bug #15642: "librbd::object_map::RefreshRequest: object map smaller than current object count: 25...
- Gotcha -- we'll get it fixed. Thx.
- 04:03 AM Bug #15642: "librbd::object_map::RefreshRequest: object map smaller than current object count: 25...
- @Jason: No. I was not able to remove the image after resizing. This is when I was hitting the error.
- 03:58 AM Bug #15642: "librbd::object_map::RefreshRequest: object map smaller than current object count: 25...
- @Rohith: Concur that we should handle this situation better -- we should force the user to disable the object map in ...
- 03:09 AM Bug #15642: "librbd::object_map::RefreshRequest: object map smaller than current object count: 25...
- Yes. agreed that there is a size limit set for object creation beyond 256000000 objects. And this works when we try t...
- 02:41 AM Bug #15642: "librbd::object_map::RefreshRequest: object map smaller than current object count: 25...
- This is because in new code base check introduced to avoid the excessively large object map creation. In this case it...
- 12:00 PM Backport #15661 (Resolved): jewel: Parent image is closed twice if error encountered while opening
- https://github.com/ceph/ceph/pull/8867
- 11:59 AM Backport #15660 (Resolved): jewel: rbd-mirror can crash if start up is interrupted
- https://github.com/ceph/ceph/pull/8866
- 09:28 AM Documentation #15212 (Resolved): rbdmap manpage is missing
04/28/2016
- 10:08 PM Feature #15656 (Resolved): Initial python APIs to support mirroring
- Needed to support integration with Cinder's replication feature.
- 08:04 PM Feature #14540: rbd disk-usage CLI command should support calculating full image usage
- *PR*: https://github.com/ceph/ceph/pull/8819
- 07:16 PM Feature #14540 (In Progress): rbd disk-usage CLI command should support calculating full image usage
- 08:03 PM Bug #15654 (Resolved): librbd: TestJournalReplay.SnapCreate hang
- Log at http://qa-proxy.ceph.com/teuthology/joshd-2016-04-27_22:40:19-rbd-wip-josh-testing---basic-mira/153344/teuthol...
- 07:56 PM Bug #15630 (Pending Backport): rbd-mirror can crash if start up is interrupted
- 07:56 PM Bug #15574 (Pending Backport): Parent image is closed twice if error encountered while opening
- 07:10 PM Bug #15572 (Fix Under Review): Errors encountered disabling object-map while flatten is in-progress
- *PR*: https://github.com/ceph/ceph/pull/8817
- 03:21 PM Backport #15649 (Resolved): jewel: possible double-free of object map invalidation request upon e...
- https://github.com/ceph/ceph/pull/8865
- 03:06 PM Feature #15648 (Resolved): librbd copy API should do a better job at sparsifying the destination ...
- Right now the entire stripe period (or object size) must be zeroed in order to avoid the write op to the destination.
- 03:03 PM Bug #15643 (Pending Backport): possible double-free of object map invalidation request upon error
- 01:38 PM Bug #15643: possible double-free of object map invalidation request upon error
- master PR: https://github.com/ceph/ceph/pull/8796
- 01:38 PM Bug #15643 (Resolved): possible double-free of object map invalidation request upon error
- 11:40 AM Bug #15642: "librbd::object_map::RefreshRequest: object map smaller than current object count: 25...
- Also attaching sample output of the commands executed:-
1)sudo rbd -p pool2 create rbd7 --size 102400000
2)sud... - 11:35 AM Bug #15642 (Resolved): "librbd::object_map::RefreshRequest: object map smaller than current objec...
- This is seen with the latest version of build ceph version 10.2.0.
Steps to reproduce:-
1) Create RBD image wit... - 08:27 AM Backport #15605 (In Progress): jewel: Potential double free of SetSnapRequest instance
04/27/2016
- 08:05 PM Bug #15574 (Fix Under Review): Parent image is closed twice if error encountered while opening
- *PR*: https://github.com/ceph/ceph/pull/8791
- 07:31 PM Bug #15574 (In Progress): Parent image is closed twice if error encountered while opening
- 07:25 PM Bug #15630 (Fix Under Review): rbd-mirror can crash if start up is interrupted
- *PR*: https://github.com/ceph/ceph/pull/8790
- 03:22 PM Bug #15630 (Resolved): rbd-mirror can crash if start up is interrupted
- If an image replay is interrupted during start up, the journaler will be left in replay mode when it's destroyed. Th...
- 06:16 PM Feature #15632 (Resolved): Expose librbd API methods to directly acquire and release the exclusiv...
- QEMU is adding support for image locking and it will be able to use these methods to lock/unlock an image and properl...
- 03:33 AM Bug #15604: RBD image mounted by command "rbd-nbd" the status is read-only.
- Pull request: https://github.com/ceph/ceph/pull/8775
04/26/2016
- 04:59 PM Tasks #15611 (In Progress): Need fsx test against the rbd-nbd device within teuthology
- 04:11 PM Tasks #15611 (Resolved): Need fsx test against the rbd-nbd device within teuthology
- FSX already supports nbd -- just need to add a test case within the rbd teuthology suite.
- 07:59 AM Backport #15605 (Resolved): jewel: Potential double free of SetSnapRequest instance
- https://github.com/ceph/ceph/pull/8803
- 07:50 AM Bug #15604: RBD image mounted by command "rbd-nbd" the status is read-only.
- I am working on this item
- 07:45 AM Bug #15604 (Resolved): RBD image mounted by command "rbd-nbd" the status is read-only.
- rbd-nbd can map and mount rbd image to /dev/nbdX successfully.
But can not write any files. The error message is "R...
04/25/2016
- 07:08 PM Subtask #14420 (Fix Under Review): New RBD notification pool object for recording mirroring errors
- *PR*: https://github.com/ceph/ceph/pull/8442
- 05:42 PM Feature #15593 (Resolved): support names in rbd default feature instead of numbering system in ce...
- I think the 'rbd default feature' in ceph.conf still uses the numbering system and the newer document has no referenc...
04/24/2016
- 11:16 AM Bug #15561: RBD image can be listed but not opened
- do you do rbd snap operations?
or you should list your steps to reproduce the problem.
04/22/2016
- 08:21 PM Bug #15577: Demote image format 1 error message on open to information message
- ...
- 08:20 PM Bug #15577 (Resolved): Demote image format 1 error message on open to information message
- Until we have an easy (zero or minimal down time) approach to migrate users from image format 1 to 2, logging an erro...
- 02:44 PM Bug #15574 (Resolved): Parent image is closed twice if error encountered while opening
- The `OpenRequest` state machine will properly close the image upon error. However, the `RefreshParentRequest` will a...
- 02:06 PM Bug #15572 (Resolved): Errors encountered disabling object-map while flatten is in-progress
- ...
- 01:38 PM Bug #15571 (Resolved): Potential double free of SetSnapRequest instance
- https://github.com/ceph/ceph/pull/8676
04/21/2016
- 11:11 AM Bug #15561 (Closed): RBD image can be listed but not opened
- root@test:/home/pair# rbd ls customer-data | grep test.cdata
test.cdata
root@test:/home/pair# rbd info data/test.cd...
04/15/2016
- 06:27 PM Bug #15522 (Resolved): Setting Image Metadata for an Image is accepting number, where it should a...
Description of problem:
Setting Image Metadata for an Image is accepting number, where it should accept only Str...- 06:26 PM Bug #15521 (Resolved): Seeing rollback and revert option for snapshot, which seems to be same and...
- Description of problem:
Seeing rollback and revert option for snapshot, which seems to be same
Version-Release nu... - 12:07 AM Bug #15457 (Resolved): Relax rbd CLI reg-ex validation for non-create/clone/import/rename operations
04/14/2016
- 08:08 PM Bug #15457 (Fix Under Review): Relax rbd CLI reg-ex validation for non-create/clone/import/rename...
- *PR*: https://github.com/ceph/ceph/pull/8606
- 04:11 PM Bug #15457 (In Progress): Relax rbd CLI reg-ex validation for non-create/clone/import/rename oper...
- 04:05 PM Bug #15500 (Resolved): qa: dynamic_features.sh races with image deletion
- http://qa-proxy.ceph.com/teuthology/joshd-2016-04-13_15:46:32-rbd-wip-josh-testing---basic-mira/127723/teuthology.log...
- 04:02 PM Bug #15102 (Resolved): librbd: exclusive lock AcquireRequest failed assert m_image_ctx.journal ==...
- 04:01 PM Bug #15487 (Resolved): Cannot successfully disable journaling with in-flight IO
- 04:01 PM Bug #15351 (Resolved): librbd: resize timed out
- 04:01 PM Bug #15471 (Resolved): Segfault while flattening image with in-flight IO
- 04:00 PM Bug #15479 (Resolved): ceph_test_rbd_mirror_image_replay: assert failed: "non_primary ^ primary_m...
- 04:00 PM Bug #15352 (Resolved): TestJournalEntries.AioWrite: failed assert(m_watch_in_progress)
- 04:00 PM Bug #15436 (Resolved): Deadlock updating dynamic features with in-flight IO
04/13/2016
- 05:02 PM Bug #14988 (Resolved): QEMU VM hangs talking to RBD via librbd
- 05:01 PM Backport #15033 (Resolved): hammer: QEMU VM hangs talking to RBD via librbd
- 05:00 PM Backport #15032 (Resolved): infernalis: QEMU VM hangs talking to RBD via librbd
- 04:36 PM Bug #15479 (Fix Under Review): ceph_test_rbd_mirror_image_replay: assert failed: "non_primary ^ p...
- *PR*: https://github.com/ceph/ceph/pull/8584
- 03:26 PM Bug #15479 (In Progress): ceph_test_rbd_mirror_image_replay: assert failed: "non_primary ^ primar...
- 05:17 AM Bug #15479: ceph_test_rbd_mirror_image_replay: assert failed: "non_primary ^ primary_mirror_uuid....
- Misread that, replacing everything with the real issue.
- 04:51 AM Bug #15479 (Resolved): ceph_test_rbd_mirror_image_replay: assert failed: "non_primary ^ primary_m...
- ceph_test_rbd_mirror_image_replay crashed in this run: http://qa-proxy.ceph.com/teuthology/joshd-2016-04-12_16:21:39-...
- 04:06 PM Bug #15487: Cannot successfully disable journaling with in-flight IO
- *PR*: https://github.com/ceph/ceph/pull/8511
- 04:06 PM Bug #15487 (Resolved): Cannot successfully disable journaling with in-flight IO
- IO deadlock and assert failures have been encountered.
04/12/2016
- 08:22 PM Bug #15471: Segfault while flattening image with in-flight IO
- The crash occurred if bench-write completed before the proxied flatten operation completed. The shut down of rbd did...
- 08:20 PM Bug #15471 (Fix Under Review): Segfault while flattening image with in-flight IO
- *PR*: https://github.com/ceph/ceph/pull/8565
- 01:08 PM Bug #15471 (Resolved): Segfault while flattening image with in-flight IO
- Start IO against a clone image:...
- 08:21 PM Bug #15456 (Fix Under Review): rbd bench-write shouldn't crash upon write failure
- *PR*: https://github.com/ceph/ceph/pull/8565
- 03:23 PM Bug #15456 (In Progress): rbd bench-write shouldn't crash upon write failure
- 06:18 PM Feature #14867: add object map verifier to RBD CLI
- https://github.com/ceph/ceph/pull/8562
- 06:14 PM Feature #14867 (Fix Under Review): add object map verifier to RBD CLI
- 07:10 AM Feature #15468 (New): Feature : cephx user management for RBD images
- Feature : cephx user management for RBD images
With current cephx user implementation[1] there is no way to achie... - 01:04 AM Bug #15351: librbd: resize timed out
- That was a subtle one -- couldn't reproduce w/ killing OSDs, hanging OSDs, etc. Needed a notify timeout during the f...
- 01:03 AM Bug #15351 (Fix Under Review): librbd: resize timed out
- *PR*: https://github.com/ceph/ceph/pull/8543
04/11/2016
- 11:30 PM Bug #15351: librbd: resize timed out
- No injection here, but could be ovh VMs being slow: http://teuthology.ovh.sepia.ceph.com/teuthology/teuthology-2016-0...
- 11:23 PM Bug #15351: librbd: resize timed out
- Pretty consistent that is happens after a "Testing filestore_inject_stall pause injection for duration 3" thrasher op...
- 10:38 PM Bug #15351: librbd: resize timed out
- also happened to flatten here: http://teuthology.ovh.sepia.ceph.com/teuthology/teuthology-2016-04-11_16:00:01-rbd-mas...
- 03:44 PM Backport #15360 (In Progress): hammer: Linking to -lrbd causes process startup times to balloon
- 02:43 PM Backport #14512 (Fix Under Review): hammer: snapshot, flatten, and rollback of cloned image leads...
- 02:32 PM Backport #14512: hammer: snapshot, flatten, and rollback of cloned image leads to data loss
- https://github.com/ceph/ceph/pull/8535
- 02:30 PM Bug #15457 (Resolved): Relax rbd CLI reg-ex validation for non-create/clone/import/rename operations
- Suggested alternative reg-ex for non-strict mode would be "^(?[^/]+)/)?([^@]+)(?:@(.+))?$"
- 01:14 PM Bug #15456 (Resolved): rbd bench-write shouldn't crash upon write failure
- If a write IO fails during bench-write, the CLI will crash due to an 'assert(0 == ret)' statement. It would be clean...
04/10/2016
04/08/2016
- 08:46 PM Bug #15259: performance: journaling results in 4X slowdown when writes are not blocked by cache
- Running against an NVMe-backed OSD:
Using bench-write on a non-journaled image:... - 07:55 PM Bug #15436 (Fix Under Review): Deadlock updating dynamic features with in-flight IO
- *PR*: https://github.com/ceph/ceph/pull/8511
- 03:49 PM Bug #15436 (Resolved): Deadlock updating dynamic features with in-flight IO
- http://qa-proxy.ceph.com/teuthology/teuthology-2016-04-06_12:01:02-rbd-jewel-distro-basic-smithi/111658/teuthology.log
- 07:55 PM Bug #15102 (Fix Under Review): librbd: exclusive lock AcquireRequest failed assert m_image_ctx.jo...
- *PR*: https://github.com/ceph/ceph/pull/8511
- 07:46 PM Bug #15102 (In Progress): librbd: exclusive lock AcquireRequest failed assert m_image_ctx.journal...
- 04:15 PM Bug #15351: librbd: resize timed out
- Not sure where it's coming from either. Seems like it would have to be one of the notifies or something watch related.
- 01:31 PM Bug #15351: librbd: resize timed out
- Having trouble figuring out where the ETIMEDOUT error came from. In theory, this error should bubble-up from the Obj...
- 12:44 PM Bug #15353 (New): librbd: disable optimizations that result in pipelining guarded writes mixed wi...
- @Josh: OK, I'm more than happy to avoid changing the IO path right before the Jewel release. :-)
- 03:51 AM Bug #15353: librbd: disable optimizations that result in pipelining guarded writes mixed with non...
- Since this wasn't caused by the recent optimizations, and we haven't seen any reports of it in the wild, I'm wonderin...
- 03:09 AM Bug #15353: librbd: disable optimizations that result in pipelining guarded writes mixed with non...
- @Josh: my assumption was that dropping the guard if we know the object exists doesn't really save us (the client) muc...
- 03:05 AM Bug #15353: librbd: disable optimizations that result in pipelining guarded writes mixed with non...
- @Josh: You will hit this issue with object map disabled when you have multiple in-flight writes to a cloned image's o...
- 02:47 AM Bug #15353: librbd: disable optimizations that result in pipelining guarded writes mixed with non...
- A fortuitous message to ceph-users makes a good point - we could also drop the guarded write when object map is used ...
- 02:44 AM Bug #15353: librbd: disable optimizations that result in pipelining guarded writes mixed with non...
- It seems like it's more probable to have the flat write overlap with guarded writes now that the object map tells us ...
- 01:42 AM Bug #15353: librbd: disable optimizations that result in pipelining guarded writes mixed with non...
- @Josh: I think I am missing something. The only times we drop the guard are during copy-ups. ...
04/07/2016
- 09:50 PM Bug #15354 (Resolved): rbd_mirror: valgrind says the 'impossible' happened - thread limit too low
- 08:42 PM Bug #15354 (Fix Under Review): rbd_mirror: valgrind says the 'impossible' happened - thread limit...
- *PR*: https://github.com/ceph/ceph/pull/8493
- 07:41 PM Bug #15354 (In Progress): rbd_mirror: valgrind says the 'impossible' happened - thread limit too low
- 09:44 PM Bug #15353 (In Progress): librbd: disable optimizations that result in pipelining guarded writes ...
- 09:40 PM Bug #15364 (Resolved): TestJournalRecorder.FutureFlush
- 09:40 PM Bug #15097 (Resolved): journal.sh fails when removing image
- 07:40 PM Bug #15097 (Fix Under Review): journal.sh fails when removing image
- *PR*: https://github.com/ceph/ceph/pull/8490
- 06:55 PM Bug #15097 (In Progress): journal.sh fails when removing image
- 09:38 PM Bug #15352 (Fix Under Review): TestJournalEntries.AioWrite: failed assert(m_watch_in_progress)
- *PR*: https://github.com/ceph/ceph/pull/8494
- 08:49 PM Bug #15352 (In Progress): TestJournalEntries.AioWrite: failed assert(m_watch_in_progress)
- 06:53 PM Bug #15233 (Need More Info): Flattening an rbd image with a running Qemu instances causes librbd ...
- @Christian: I just occurred to me that the fix for the original issue, while flagged as resolved, is still forthcomin...
- 02:40 PM Bug #15233 (In Progress): Flattening an rbd image with a running Qemu instances causes librbd wor...
- 02:42 PM Backport #15415 (Duplicate): infernalis: Flattening an rbd image with a running Qemu instances ca...
- 02:42 PM Backport #15414 (Duplicate): hammer: Flattening an rbd image with a running Qemu instances causes...
04/06/2016
- 09:39 PM Bug #15364 (Fix Under Review): TestJournalRecorder.FutureFlush
- *PR*: https://github.com/ceph/ceph/pull/8477
- 03:30 PM Bug #15364 (In Progress): TestJournalRecorder.FutureFlush
- 03:37 PM Bug #15233: Flattening an rbd image with a running Qemu instances causes librbd worker threads to...
- 12:06 PM Bug #15233: Flattening an rbd image with a running Qemu instances causes librbd worker threads to...
- Thanks -- so what client applications does this log output include? Does this log combine the rbd CLI and qemu? What...
- 06:22 AM Bug #15233: Flattening an rbd image with a running Qemu instances causes librbd worker threads to...
- It seems the client log went missing. Here it is (again). Ah it's too big. Redmine says max. size 73.4MiB, but your n...
- 03:01 PM Backport #15129 (Resolved): hammer: Protect against excessively large object map sizes
- 02:34 PM Backport #15033: hammer: QEMU VM hangs talking to RBD via librbd
- @Stefan: are you able to recreate your lockup and provide "debug rbd=20"-level logging?
- 02:32 PM Backport #15033: hammer: QEMU VM hangs talking to RBD via librbd
- but this backport was just merged into the hammer branch ;-)
- 02:31 PM Backport #15033: hammer: QEMU VM hangs talking to RBD via librbd
- hammer-backports is not a branch that is for usage, it is solely for test purposes
- 01:50 PM Backport #15033: hammer: QEMU VM hangs talking to RBD via librbd
- Hammer Backports was force pushed.
last time i tested it there was the following commit:
commit 7d6737ea82f23f5... - 12:23 PM Backport #15033: hammer: QEMU VM hangs talking to RBD via librbd
- @Stefan: are you sure this fix was in hammer-backports when you ran your test or are you saying you merged it manuall...
- 02:13 PM Bug #15404 (Can't reproduce): KVM: terminate called after throwing an instance of 'ceph::buffer::...
- KVM instances throwing "terminate called after throwing an instance of 'ceph::buffer::end_of_buffer'" after adding ov...
Also available in: Atom